AI-Assisted Software Development

148 stories · publishing
Updated Sep 2, 2026
AIU research department

PublishingThis department publishes coverage today — every story links out to its own source.

Share this departmentXLinkedInEmail

The lead

Sep 2, 2026

AI-Assisted Software DevelopmentAnthropicmajorSep 1, 2026

Anthropic released Claude Fable 5.1, priced about 25% below Fable 5 for typical work

Anthropic released Fable 5.1 and Mythos 5.1 on September 1 — the same underlying model behind two different safeguard settings, with Fable generally available and Mythos limited to trusted-access programs.

What it means for your work

The change that shows up on a bill is the cache-read price, which is where long agentic runs spend; Devin's team said it is what finally made a Fable-class model economical for their code review.

✓ verified · anthropic.com · added todayRead it at anthropic.com

What today means

Our read on the items that move something. The reporting is everyone's; this part is ours.

  1. The last mile of a pull request — rerunning checks, clearing conflicts, answering review notes — is the part that actually eats an afternoon.VS Code 1.136 adds an agent that works a pull request until it is ready to mergeThe AI-Assisted Software Development beat · Visual Studio Code

  2. Third Flash release in six weeks at the same rate card — the cheap tier is where most production traffic actually runs, and working harder per task is a cost change even when the price is not.Google's Gemini 3.8 Flash holds the old price and adds a cybersecurity-only siblingThe AI-Assisted Software Development beat · Google DeepMind

  3. Changing model or provider is normally an application rewrite; a translating proxy turns it into a routing rule you can measure both sides of.NVIDIA opened a proxy that lets one application speak both the OpenAI and Anthropic APIsThe AI-Assisted Software Development beat · NVIDIA

4 new findings here overnight. 142 here in all.

What this department watches

The charge

This department watches how software actually gets built once an AI assistant is inside the loop: the coding models and agents, the editors and command-line tools that host them, and the review, testing and deployment practice teams are rewriting to keep pace. It follows what holds up in production over what reads well in a launch post — the interesting evidence is what a team reports six months after adoption, not on the day of it.

  • Coding models and agent frameworks, and what they measurably do
  • Editors, command-line tools, and the toolchain around them
  • Review, testing and correctness practice under AI authorship
  • Deployment, reliability, and incident practice
  • Security of AI-written and AI-assisted code
  • What engineering teams report after adoption, not before

The questions it keeps open

6 standing questions

These are the open questions this department works. They are questions, not conclusions — what comes back gets published with its sources attached.

  • When AI generates a large share of a change, what does responsible code review look like — and what does the reviewer have to read line by line versus verify by behaviour?
  • Which parts of the toolchain are consolidating and which are still churning, so a builder knows what is safe to build a workflow on?
  • What failure modes do AI-assisted codebases actually hit in production — and are they new failures or old ones arriving faster?
  • How do testing and debugging practice change when the code's author can't explain its own reasoning?
  • What does a hiring manager or a paying client now treat as proof of engineering capability, when “wrote the code” no longer distinguishes anyone?
  • Where does agent autonomy stop paying — the point past which supervising the work costs more than doing it?

How the research runs

The process
1 · Read the real sources

AIU's research agents work from a published register of trusted sources rather than from memory — the register itself is open, so you can see what is being read and how often it is checked.

The source register →

2 · Carry the method and the source

A finding is only useful if you can check it. Every item carries the source it came from and the date that source published, and links straight out to the original — so a claim can be verified rather than taken on trust.

The research archive →

3 · Feed the coursework

What this department reads is what keeps the coursework current — and the coursework decides what is worth watching next. The two are meant to inform each other, which is why the curriculum below is on this page rather than somewhere else.

The curriculum behind it

3 courses on aiuni.tech
The major

AI-Assisted Software Development

Build software with the current generation of AI coding tools — the self-directed builder's alternative to a CS degree.

Follow this department

Two doors

This department is publishing now — read the edition, or put your name down to be told when its research goes further.

Request early access to this department's research Read the AI-Assisted Software Development edition

Behind this department

Everything below is Intel's one record, filtered to this department — the same pages Research opens, showing only this slice. Each one says so on arrival and links back to everything.
Sources for this departmentOnly the sources it draws onResearch behind this departmentEvery finding it selected, in fullThis department on the mapIts slice of the knowledge graph
All ten departmentsOne department per page · charter and process, never invented findings